摘要
提出了一种航拍机场跑道的快速识别方法。采用二维最大熵阈值方法分割出目标区域,利用Kirsch算子提取主要物体的边界轮廓,然后基于数学形态学的开、闭运算消减云层等物对目标边界的干扰,断开细线连接,削弱狭窄的部分,再应用细化剪枝剔除毛刺和各段孤立的短骨架线;利用Hough变换搜索出平行直线对作为机场跑道的候选区域;利用跑道区域的灰度特征对候选区域进行验证,去除虚假目标。实验结果表明:该方法具有较高的目标识别率和计算实时性,抗噪性强,能够消除背景中诸如云朵等物体的干扰。
Presented a fast method to detect airport in aerial image. First, the two-dimensional maximum entropy threshold rnethod was adopted to segment the object area, and Kirsch operator was used to detect the main edges of the objects in the image. Then the disturbance of some background objects such as cloud was eliminated. Also thread linking and the narrow part were disconnected and weakened accordingly, which were based on the mathematical morphology of the opening and closing operation. Further the burr and every ＇isolated short skeleton were removed by thinning and pruning. And then, the parallel straight lines detected by Hough transform were ennsidered as candidate regions. At last, the candidate regions were verified by the gray characters of runway area so as to wipe off the false targets. Experiments indicate that the method has a high recognition rate and calculation of real- time, and it is robust to the background interference.
出处
《计算机技术与发展》
2008年第7期193-196,共4页
Computer Technology and Development
基金
重庆自然科学基金项目(CSTC
2005BB2207)
关键词
二维最大熵阈值
数学形态学
细化剪枝
机场检测
2 - D maximum entropy threshold
mathematical morphology
thinning and pruning
airport detection
